migrating from CVS

[Jos Backus]

    Hi Martin,

On Mon, Mar 06, 2006 at 11:18:24AM +1100, Martin Pool wrote:
> 
> On 2006-03-06 Mon, at 8:17 AM, Jos Backus wrote:
> 
> Hi Jos, nice to see you again.

Thank you, and good to see you too. To me it's confidence-inspiring to see
your name associated with Bazaar-NG.

> >Some questions:
> >
> >As I understand it, the Bazaar-NG model is that one has a checkout  
> >(tree) for
> >every branch, whereas with CVS there can be one tree but checked- 
> >out files can
> >be on different branches. We use this feature to introduce new  
> >software in an
> >otherwise stable tree.
> >
> >Is my understanding of the above correct as far as Bazaar-NG goes?
> 
> This is the default case, but we are just finalizing changes which  
> allow other modes of operation.  You can (in bzr.dev) have a  
> repository holding multiple branches, but without any working trees  
> present.  Robert is now preparing to merge John's work on checkouts,  
> which allows you (as with CVS) to have a directory holding a working  
> copy but no other metadata.

Both these other modes sound useful, too. As I understand it, the default mode
is common to Subversion and Bazaar-NG so no differences there. With both
systems, all files in a tree by definition belong to a single branch. This is
a fundamental difference with CVS, where branching happens at the file level
(as in the examples I gave). Am I getting it?

> >Say I have a CVS checkout with 1000 files, all of which are on the  
> >trunk but
> >one file lives on a branch. Does this mean that in Bazaar-NG land I  
> >would have
> >two almost identical trees except for the one file?
> 
> Yes, that's correct.  Or you can keep them in one repository, which  
> will mean the data is stored approximately as compactly as in CVS.
 
So this means that the source of the checked-out (a.k.a. working) trees, each
mapping to a single branch, is a single (shared) repository? I'm assuming the
sharing pertains to file/directory (meta-)data between the branches that the
working trees are an instantiation of?

Thanks for taking the time to answer my questions, Martin.

> -- 
> Martin

Cheers,
-- 
Jos Backus
jos at catnook.com